1 Introduction
1.2 Study methane dynamics in marine sediments using proxies
1.2.2 Methane derived authigenic carbonates (MDAC)
AOM is also closely related to carbonate formation at shallow depths beneath the seafloor, as
one product of the AOM process is bicarbonate. Bicarbonate can increase alkalinity and form
authigenic carbonates as aragonite, calcite, and dolomite depending on the cation (i.e. Mg
2+,
Ca
2+) concentrations. These three are the main carbonate phases associated with methane
seeps; the mineralogy correlated to the predominant pore water cation composition at the
depth of formation(Burton, 1993; Ferrell and Aharon, 1994; Bohrmann et al., 1998) and can
also provide some insight into the precipitating environment. For example, the aragonites are
believed to be formed at high sulfate concentrations while the high sulfate would inhibit
calcite formation (Bohrmann et al., 1998; Aloisi et al., 2000). Seep carbonates serve as a good
chemical archive for methane seepage. They are characterized by negative
13C values often
below -30 ‰ (Peckmann and Thiel, 2004), which are inherited from
13C-depleted methane,
reflecting their light carbon sources (Claypool and Kaplan, 1974; Whiticar, 1999). The
18O of
the seep carbonates can also reveal information about the precipitating environment. In the
sediments containing gas hydrate, hydrate dissociation would produce an elevated
18O
signature in the seep carbonate (Bohrmann et al., 1998; Aloisi et al., 2000; Bohrmann et al.,
2002). Seep carbonate serves as suitable housing for the AOM microbial communities, lipid
biomarkers extracted from the carbonates along with other biogeochemical information can
shed some light on the methane seepage at the time of carbonate precipitation.
8 1.2.3 Foraminifera
Foraminifera are single-cell protists with calcified shells or tests. They have pseudopods, fine strands of cytoplasm, and live in the marine domain (Sen Gupta, 2003). Foraminifera are abundant as fossils for the last 540 Ma. Foraminifera can be found in all marine settings, from the cold seeps to hot vent. Some of them live in the water column floating freely, these are known as planktonic foraminifera. The others live on the seafloor (epibenthic) or in the sediments pore space (infaunal), these are known as benthic foraminifera. Their species assemblages, especially the benthic ones, can be very particular and provide information about the environment they live in (Horton, 1999; Todo et al., 2005). Depending on the species, foraminifera develop different chambers of their calcified shells (tests) when they grow. The shell can consist of calcite or aragonite and/or organic compounds (Bentov and Erez, 2006; de Nooijer et al., 2014). Because foraminifera are everywhere in the marine realm, they are one of the most essential biological proxies to study the paleoenvironment (Armstrong and Brasier, 2005).
Foraminifera have been used as geochemical proxies to reconstruct the paleo seepage at different locations such as Cascadian margin Pacific Ocean (Rathburn, 2000; Rathburn et al., 2003; Hill et al., 2003; Hill et al., 2004; Bernhard et al., 2010), Blake ridge Atlantic Ocean (Panieri and Sen Gupta, 2008), the Mediterranean (Panieri, 2006) and Vestnesa Ridge (Schneider et al., 2018). Foraminifera have also been studied in modern seep settings to explore their biological response to methane (Bernhard et al., 2010; Bernhard and Panieri, 2018), as well as the origination of changes in stable isotope composition of their tests (Rathburn et al., 2003; Torres et al., 2003b; Panieri et al., 2009).
Figure 4. Modified from Schneider et al., 2019
9
The
13C of planktonic foraminifera in a standard marine setting without methane discharge range between -0.5 to 0.5 ‰ in the Barents sea (Knies and Stein, 1998). The
13C of benthic foraminifera, such as C. neoteretis, in a typical marine environment absent of methane seepage range between 0 to -1.15 ‰ (Wollenburg et al., 2001). More depleted carbon isotopic excursions in modern and fossil foraminiferal tests have been interpreted as a result of methane influence, and have been used to imply the strong
13C-depletions in the tests are associated with
13C-depleted methane. (Kennett, 2000; Hill et al., 2003; Panieri et al., 2009;
Martin et al., 2010; Panieri et al., 2014).
The variations and extent of negative excursion in δ
13C of foraminiferal tests at the seep sites are likely the combined outcome of four aspects (Figure 4). Species-specific vital effects, nutrition sources as the ingestion of
13C-depleted AOM microbes (archaea or bacteria) the foraminifera may feed on (Rathburn, 2000; Panieri, 2006; Bernhard and Panieri, 2018), calcification using a
13C-depleted DIC (primary), and diagenetic alteration of the foraminiferal tests.
Vital effect based on the species differences can account for 1-2 ‰ of the carbon isotopic values of the foraminiferal shells (Urey et al., 1951; McCorkle et al., 1990; Mackensen et al., 2006). Foraminifera are attracted to rich organic and microbial food, and some foraminifera were found to prefer the seep-associated microbes (Panieri et al., 2009; Martin et al., 2010).
Nonetheless, the nutrition sources can contribute only up to 5-6 ‰ of the negative δ
13C values (Hill et al., 2003).
Primary biomineralization, foraminifera developing their shells with the depleted DIC produced by AOM is suggested to be another factor that may contribute to the negative values of foraminiferal tests. Some research show that living Cibicides wullerstorffi’s test become depleted in a culturing experiment with methane-derived DIC, implying the test biomineralization takes place in the presence of methane-derived DIC (Wollenburg et al., 2015). Others have doubts that biomineralization can not occur during seepage as the equilibrium between foraminifera and porewater DIC is not reached (Rathburn et al., 2003;
Torres et al., 2003b; Herguera et al., 2014).
The most substantial influence on the negative carbon excursion of foraminiferal test is the
diagenetic alteration of the tests under the methane influence (Rathburn, 2000; Rathburn et
10
al., 2003; Hill et al., 2003; Torres et al., 2003b; Panieri, 2006; Panieri et al., 2009; Panieri et al., 2014; Panieri et al., 2017b; Schneider et al., 2017; Schneider et al., 2018). The
13C-depleted bicarbonate produced during AOM can precipitate not only as authigenic carbonates concretions (Aloisi et al., 2002; Reitner et al., 2005) but also on the foraminiferal shells. Once dead, both benthic and planktonic species can record the
13C signature from the AOM process by acting as a ‘template’ for authigenic carbonate to precipitate coating layers on (Panieri et al., 2016; Panieri et al., 2017b; Schneider et al., 2017). Such coating carbonate precipitation at the SMTZ cumulatively added a second or third layer of
13C-depleted carbon to the foraminiferal tests is termed as diagenetic alteration (Schneider et al., 2017). These coating layers usually exhibit different states of shell preservation and very depleted
13C values up to -20 ‰ (Panieri et al., 2016; Panieri et al., 2017b; Schneider et al., 2017). Both C. neoteretis and
N. pachyderma are excellent templates for the authigenic carbonate formation (Panieri et al.,2017). As a result of multiple coating layers, the diagenetic alteration of foraminiferal tests can cause a much more profound depleted
13C signal (Torres et al., 2003b; Hill et al., 2004;
Panieri et al., 2009; Martin et al., 2010; Schneider et al., 2017).
1.2.4 Sediment properties
The sediments experienced active methane seepage collect diagenetic overprints as a result of the AOM process. The products of AOM, bicarbonate (HCO
3-) is consumed in ambient DIC, carbonate precipitation, and foraminifera shell. The other product of AOM is hydrogen sulfide (HS
-), which also increase alkalinity, and can react with iron (II) in the pore water and yield paramagnetic pyrite (FeS
2) (Canfield and Berner, 1987; Peckmann et al., 2001; Riedinger et al., 2006; Dewangan et al., 2013). At the same time, metastable greigite (Fe
3S
4) can form during the pyritization process as a precursor to pyrite (Hunger and Benning, 2007). Both the paramagnetic authigenic pyrite and ferromagnetic greigite can then reduce the magnetic susceptibility of the original sediment magnetic properties.
1.2.5 Others
Other common proxies such as barite formation at the base of SMTZ, dense benthic
macrofaunal communities, and AOM biofilm appearance are also used for tracing the
methane seepage in combination with the previously discussed proxies.
11
Briefly, in the sulfate depletion zone below the SMTZ, barite is destabilized and dissolved.
Barite can re-precipitate above the SMTZ as sulfate become available again, as the fluid migrates upwards, the barite formation at the base of SMTZ is known as ‘barite front’ (Torres et al., 1996; Dickens, 2001; Torres et al., 2003a; Solomon and Kastner, 2012). The barite front can be detected with geochemical analysis of barium concentration in both sediment and pore water (Kasten et al., 2012), a simplified way of identifying barite front is from the Ba/Ti ratio or Ba counts of XRF scan of the sediment cores (Sauer et al., 2016). Discovery of a discrete shell bed dominated by Vesicomydae (Phreagena, Isorropodon) at Vestnesa Ridge was interpreted as a high flux seepage episode (Ambrose et al., 2015; Schneider et al., 2018). The seepage not only supports elevated macrofauna biomass, but high methane flux can also support elevated AOM microbial biomass (Yao et al., 2019). In unusual cases, the biomass accumulated so much that form biofilm, which is rarely observed (Briggs et al., 2011; Gründger et al., 2019), but the presence of biofilm is very reliable and serve as a direct piece of evidence for AOM and thus methane seepage.
1.3 Study areas
1.3.1 Vestnesa Ridge
Vestnesa Ridge (79 °N, 5-7 °E, Figure 5), northwest of Svalbard, is one of the northernmost hydrate reservoirs. The water depth is 1200 to 1300 meters, and the ridge is a 100 km long sediment drift on the eastern Fram Strait. Fram Strait was the only deep-water gate to the Arctic Ocean, and it was opened during the late Oligocene to Miocene. The final opening of Fram Strait during the late Miocene (Jakobsson et al., 2007; Knies et al., 2014) led to the development of over 2 km thick sediment accumulation at the eastern segment of the ridge.
Moreover, the shallow stratigraphy consist of contourite, turbidite and hemipelagic sediments
have been worked by the ocean bottom currents (Howe et al., 2008).
12 Figure 5. An overview map of the Arctic Ocean with the location of two study areas: Vestnesa Ridge and Storfjordrenna, image section from IBCAO3.0 (Jakobsson et al., 2012).
Since the discovery of pockmarks, the semi-circular seafloor depressions, by Vogt et al. (1994,
1999) in Vestnesa Ridge, the area was mapped thoroughly and well-studied by variously
geophysical approaches. Pockmarks are formed under vigorous gas and fluid seepage in
unconsolidated sediments (Judd and Hovland, 2007). In addition to the pockmarks as
morphological evidence, the eastern segment of Vestnesa Ridge is also characterized by up
to 900 m high gas bubble streams (or termed as hydroacoustic flares) in the water column
(Smith et al., 2014; Panieri et al., 2017a) (Figure 6) and acoustic chimneys in the sediments
from the seismic data as gas migration pathways. It is suggested the methane seepage in
Vestnesa was driven by a rare bottom-up mechanism, where heat from the nearby mid-ocean
ridge system perturbs the gas hydrate stability after the investigation of the local seismic (Bünz
et al., 2012). The fluid and gas migration from deep hydrocarbon reservoirs toward the
seafloor has occurred since the early Pleistocene (Knies et al., 2018).
13 Figure 6. Seafloor bathymetry of the eastern section of Vestnesa Ridge, with gas flares emitting to the water column(above); gas migration pathway and BSR in the seismic data (left below); Seismic profile outlining vertical gas migration pathway and acoustic anomalies beneath Lomvi pockmark (right below). Figures from Panieri et al., 2017a.
Vestnesa Ridge hosts a very complex fluid system from the deep hydrocarbon reservoirs
(Petersen et al., 2010; Bünz et al., 2012; Plaza-Faverola et al., 2015; Panieri et al., 2017a). The
ridge actively releases methane from seafloor only along the eastern segment of the ridge
(Bünz et al., 2012; Plaza-Faverola et al., 2015; Panieri et al., 2017a). It has also been suggested
that tectonic stress field controls subsurface faulting and rifting which results in the observed
seafloor methane seepage (Plaza-Faverola et al., 2015). Modeling indicated that the onset of
the hydrocarbon discharge was the result of the rapid burial of hydrocarbon source after the
14
onset of Northern Hemisphere glaciations which lead to an increased sediment deposition (Knies et al., 2018). It is this hydrocarbon system from 2.7 Ma ago, that predominantly controls the deep thermogenic methane fluxes and seepage dynamics in Vestnesa Ridge over geological times (Knies et al., 2018).
Recent studies in the area on the
13C of foraminifera and methane derived authigenic carbonates correlated the past and ongoing methane seepage and subseafloor methane cycling to the glacio-isostatic adjustment (Schneider et al., 2018; Himmler et al., 2019). New evidence from U-Th dating of methane derived authigenic carbonates also shown that the seepage timing is linked with the wax and wane of the ice sheet (Himmler et al., 2019). Glacio-isostatic adjustments may have triggered the re-activation of tectonic faulting at Vestnesa Ridge and induced the fluid migration pathway for methane transport.
1.3.2 Storfjordrenna Gas hydrate mounds
Storfjordrenna or Storfjorden Trough (76 °N, 15- 16 °E), is located ~50 km south of Svalbard
(Figure 5), and the water depth of around 380-400 m (Serov et al., 2017). Storfjorden trough
is the second-largest trough in the western Barents Sea and is strongly affected by the ice
sheet dynamics. The trough was developed by a dynamic ice stream draining substantial
portions of the Barents Sea Ice sheet (BSIS) during the glaciation. Very different from Vestnesa
Ridge, Storfjordrenna represents a shallow-water gas hydrate system, which can be directly
affected by the bottom water warming and pressure changes induced by ice sheet
retreatment (Serov et al., 2017). Indeed, ice sheet modeling suggests that Storfjordrenna was
covered by grounded ice up to 2 km in thickness from 33 Ka to 19 Ka BP (Patton et al., 2017)
After the deglaciation, relaxation of the underlying lithosphere leads to the glacio-isostatic
adjustments which are still happening today (Auriac et al., 2016).
15 Figure 7. Seafloor morphology, gas chimneys in the seismic, and gas flares emitting to the water column in the Storfjordrenna gas hydrate mounds or gas hydrate pingo (GHP) as denoted in the figure. Figure from Serov et al. (2017).
Several mounds in Storfjordrenna were discovered during 2015 research cruises. These mounds with gas hydrate underneath were named gas hydrate mounds (GHMs) or gas hydrate pingos (GHP, Figure 7) were around 10 m in height and 500 m in width. They feature gas flares above the mounds, and hydrates were recovered from several of them (Hong et al., 2017;
Serov et al., 2017). Earlier investigations and modeling suggest that the methane seepage in
Storfjordrenna was linked with the ice sheet dynamic as the area was in the glaciated area,
and the shallow water depth could be changed due to glacial isostatic adjustment. The gas
hydrate stability zone (GHSZ) thickness change as the ice sheet advanced and retreated (Serov
16
et al., 2017, Figure 8 below).
Figure 8. modeling of the relationship between the gas hydrate stability zone and ice sheet coverage in 37,000 year time frame. Figure from Serov et al. (2017).
Seismic data revealed sub-vertical amplitude masking zones beneath the GHMs as the fluid migration pathways (Waage et al., 2019). The upper Paleocene-Eocene and Pliocene-Pleistocene sedimentary rocks offer high-permeability zones for the gas and fluid migration.
Waage et al (2019) observed a clear relationship between the thermogenic methane system
in Storfjordrenna GHMs and the regional fault system, which could potentially establish a
typical scenario of fault-controlled methane migration across the whole Svalbard- Barents Sea
margin.
17
1.4 Methodology
1.4.1 Core collection and sediment properties
In this thesis, the sediment cores were collected from three CAGE cruises CAGE 15-2, CAGE 15-6, CAGE 16-5 (all cruise reports are accessible from the CAGE website)and an NGU cruise P1606. The gravity cores were collected in Storfjordrenna gas hydrate mounds 3 (active) and 5(inactive). Upon recovery of the cores, they were cut into 1m sections, split longitudinally into working and archive halves. Subsamples for stable isotope analyses of foraminifera and authigenic carbonates, lipid biomarker and radiocarbon dating were sampled onboard in the working half. The archive half was stored at 4 °C for sedimentary XRF and MSCL scanning onshore at UiT Geolab. The multi-cores from Vestnesa Ridge were collected by the multicorer system. The system was equipped with a MISO (Multidisciplinary Instrumentation in Support of Oceanography, Woods Hole Oceanographic Institution) towcam. Each deployment collect six paralleled cores at most (Panieri et al., 2015;Panieri et al., 2017a). Among the six cores, one was assigned for porewater analyses, and the two adjacent cores were subsampled for lipid biomarker and headspace gas analyses, respectively. A push core was collected during cruise P1606 using the remotely operated vehicle (ROV) Ægir with a video survey (Yao et al., 2019).
Sedimentary property measurements such as magnetic susceptibility, XRF, and X-ray scanning were done at UiT Geolab on the archive half. Magnetic susceptibility was acquired in 1 cm interval using a GeoTek Multi-Sensor Core Logger (MSCL) on all the gravity cores (except 1520GC). X-ray fluorescence (XRF) element-geochemical data were attained with an Avaatech XRF Core Scanner at 1 cm resolution. All the archived halves of the sediment cores and archiving multicores were scanned with a GEOTEK X-ray core imaging system (MSCL-XCT 3.0), using an X-ray intensity of 120 kV and a measuring resolution of 1 cm.
1.4.2 Stable isotope of foraminifera tests and carbonates
Foraminiferal and carbonates
13C and
18O analysis were done at the stable isotope
laboratory at UiT using a Thermo Scientific MAT253 Isotope Ratio Mass Spectrometer (IRMS)
coupled to a Gasbench II. Studies in this thesis were done on two benthic: Cassidulina
neoteretis (Seidenkrants 1995), Melonis barleeanus (Williamson 1858), and one planktonic(Neogloboquadrina pachyderma (Ehrenberg 1861)) foraminiferal species. These species were
18
picked due to their high abundance in the study areas after an overview survey of all the foraminiferal samples. Carbonate nodules/crusts were sieved out, grounded into homogeneous powder using a quartz mortar. Foraminiferal tests or carbonate powder were placed in specific vials and flushed with helium gas before five drops of water-free H
3PO
4were added manually. After equilibration (>3 hours at 50°C), the samples were analyzed on the IRMS. Normalization to the
Vienna Pee Dee Belemnite (VPDB) for carbon and oxygen isotopes were done using in-house standards. Analytical precision was better than 0.07 ‰ for δ
13C and 0.08 ‰ for δ
18O by measuring the certified standard NBS-19 repeatedly in the sequence queue.
Carbonate nodules/crusts were grounded and analyzed on a Bruker D8 Advanced diffractometer (Cu Ka radiation in 3-75 ° 2θ range) at NGU (Sauer et al., 2017) for mineralogy.
The quantification of the carbonate mineralogical composition phases was modeled using the Rietveld algorithm-based code Topas-4 by Bruker. The correction of the spectrum was made on the main quartz peak, and the displacement of calcite d104 was used to estimate the amount of MgCO
3mole percentage (Goldsmith et al., 1958).
1.4.3 Lipid biomarkers of sediment and carbonates
Sediment lipid biomarkers were extracted and analyzed according to previously reported protocols (Elvert et al., 2003). Briefly, a total lipid extract (TLE) was attained by ultrasonication of
∼ 20 g wet sediment samples in four steps using solvents with decreasing polarity:dichloromethane (DCM) / methanol (MeOH) 1 : 2; DCM/MeOH 2 : 1; and only DCM for the last two steps. Carbonate lipid biomarkers were extracted similarly but were washed and acidified by 37% HCl before the TLE extraction.
The TLE was saponified with NaOH, the resulting neutral fraction was extracted with hexane prior to methylation to produce fatty acid methyl esters (FAMEs) for chromatographic analysis.
The positions of the double bonds in FAMEs were determined by analyzing the corresponding dimethyl–disulfide adducts (DMDS) (Nichols et al., 1986; Moss and Lambert-Fair, 1989). With pipette column chromatography, the neutral fraction was further separated by solvents with increasing polarity into hydrocarbons, ketones, and alcohols. The alcohol fraction was derivatized to form trimethylsilyl (TMS) adducts for analysis.
The individual lipid compound was analyzed using gas chromatography (GC) (Thermo
Scientific TRACE™ Ultra), equipped with a capillary column (Rxi-5ms, 50 m, 0.2mmID, 0.33 µm
df), helium gas was used as a carrier gas at a constant flow rate of 1mL min
−1. The initial oven
19
temperature was set to be 50 °C, held for 2 min. Increased to 140 °C at a rate of 10 °C min
−1,
and held for 1 min. Further increased to 300 °C at 4 °Cmin
−1. The final hold time was 63 min to
analyze FAMEs and 160 min for the neutral hydrocarbon and alcohol fractions to analyze
higher boiling points lipids. Concentrations of the lipids were determined by flame-ionization
detection (FID) against internal standards. Unknown compounds were identified with a
quadrupole mass spectrometry (QMS) unit (Thermo Scientific DSQ II) at the chromatography
periphery. Using the same temperature program, compound-specific stable carbon isotope
ratios were determined using a magnetic sector isotope ratio mass spectrometry (Thermo
Scientific Delta V Advantage) coupled to a GC setup the same as the above-mentioned
specification. δ
13C values are reported with an analytical error of ±1‰.
20
2 List of scientific contributions
o
List of first author scientific paper and manuscripts:
Paper 1:
Yao, H., Hong, W. L., Panieri, G., Sauer, S., Torres, M. E., Lehmann, M. F., Gründger, F., and Niemann, H.: Fracture-controlled fluid transport supports microbial methane-oxidizing communities at Vestnesa Ridge, Biogeosciences, 16, 2221-2232, 10.5194/bg-16-2221-2019, 2019
Paper 2:
Yao, H., Niemann, H., and Panieri, G.: Multi-proxy approach to unravel methane emission history of an Arctic cold seep (submitted to Quaternary Science Reviews)
Paper 3:
Yao, H., Panieri, G., Lehmann, M., Himmer, T., and Niemann, H.: Biomarker and isotopic composition of seep carbonates record environmental conditions in two Arctic methane seeps (to be submitted to Deep Sea Research)
o